The question on how the stock price reacts to the dividend announcement has been widely studied across different countries. Evidence from the Russian market is scarce while existing studies claim counterintuitive negative reaction of stock price to the positive dividend surprise. We show that one have to be accurate with both methodology and data source while studying that question. First, the simple rank or t-tests for event studies do not account for the panel structure that seem to be important. Second, the dividend surprises may be measured inaccurately. We present evidence that results depend on the data source used, Bloomberg or I/B/E/S Thomson Reuters, where the latter may be less reliable for some of the series. We use panel data methods with Bloomberg data and show that the results for the Russian market is standard positive association of the stock price reaction to the dividend surprise.